Paste one prompt into your AI, answer four questions, and you'll have a click-based sending segment built inside beehiiv.
You don't need to understand every branch of a Base Sending Segment to build one anymore. You need one prompt and about ten minutes.
Here's the thing.
A few weeks ago I told you your engaged segment is probably full of ghosts, subscribers who qualify because a pixel fired, not because a person showed up.
The fix is a segment built on clicks.
But building it by hand means four branches, four windows, and a lot of room to get one of them wrong.
So I wrote the instructions down for your AI instead.
It's a guide on our site with a section written just for the assistant.
Point Claude or ChatGPT at it, with the beehiiv MCP connected, and it walks you through every decision one question at a time, checks your actual sending history, and builds the segment in your account when you say go.
Copy This Into Your AI
Make sure the beehiiv MCP is connected to your AI first. Then paste this in exactly as written:
Read this guide and follow the "Instructions for the AI assistant" section exactly: https://www.audiencebridge.io/p/creating-a-base-sending-segment-using-the-beehiiv-mcp
Then create a Base Sending Segment for my beehiiv publication. Ask me one question at a time, using tappable options where you can. Don't create anything until I've confirmed the final settings.That last line matters. Nothing gets built in your account until you've seen the final settings and said yes.
It Asks Four Questions, and Checks Your Answers
I recorded myself running it on one of our publications so you can see exactly what happens.
The first thing it asks is how often you send. I said daily.
It went and looked at what we'd actually sent and came back with 6.5 a week, because we'd skipped a day or so around a holiday.
Close enough.
But that's the point: it checks your cadence against your real sending history before it builds anything on top of it.
Cadence matters because every window in the segment counts sends, not calendar days.
Ninety days means something completely different when you send twice a day than when you send once a week.
The guide does the conversion for you.
Then it asks four questions, one window each. Here's what I picked for a daily send:
Window | The question it's answering | My setting |
|---|---|---|
Click | How many sends in a row without a click before a past clicker drops out? | 90 sends (~90 days) |
Open | How far back does an open still count, for someone who's clicked before? | 60 sends (~60 days) |
Grace | How long can a new subscriber stay on opens alone while you figure out if they're real? | 120 sends (~120 days) |
New subscriber | How long do brand-new subscribers get every email, no activity required? | 15 sends (~15 days) |
I give the grace window a long leash on purpose.
New subscribers open a lot early, and some of those opens are bots. A longer grace window gives the humans time to prove it with a click.
If you're not sure on any of them, tap "not sure." It'll recommend a starting value based on your cadence.
That's what I did for the new subscriber window, and it suggested 15 sends.
One rule the guide enforces: the grace window has to be longer than the new subscriber window.
If it isn't, the grace branch never does anything.
What It Built for Me
Once I confirmed the settings, it created the segment in beehiiv and named it with the windows baked in: BSS - 90/60/120/15.
That way you can tell at a glance what any version is running.
Don't panic if the count shows nothing at first. beehiiv takes a few minutes to process a new segment.
I refreshed a couple of times, and it landed at 83,839 subscribers.
Those are the people getting our daily send now.
Everyone in there has either clicked recently, opened recently with a click somewhere in their history, or is new enough that they haven't had a fair chance yet.
Now, 83,839 on its own doesn't tell you much. The number that matters is what share of your list it is.
Out of 289,919, that's 29% of the list we were sending to that has shown a human signal.
Everyone outside it was costing us placement without giving anything back.
Run yours and do the same math.
If the share comes in lower than you expected, that's not the segment being too strict.
That's how much of your sending was going to people who weren't there.
I'll be straight about who this fits.
Everything I run is built around clicks. That's how we make money.
Sponsors pay when their ads get clicked, and we drive readers to our sites.
So the segment I send to is built on the only signal a machine doesn't fake at scale for you, which is a person clicking something.
If your newsletter is pure content and you're not asking for clicks, this version will look harsh, because your readers never had a reason to leave a click behind.
I'll put together an opens-based version for that down the road.
If clicks are how you get paid, though, this is the segment you should be sending to.
Build It, Then Watch What Happens
This isn't something you have to switch over to all at once.
Build the segment first.
Look at the count, work out the share, and compare it to whatever you're sending to now.
Then test it.
Send to the BSS for a couple of weeks and watch your click rate and your Postmaster data.
If placement tightens up, keep going.
If you want to adjust a window, run the prompt again with different answers. The name tells you which version is which.
And if you get stuck anywhere along the way, hit reply.
I read them.
Enjoy testing the prompt,
Chris Miquel
P.S. A Base Sending Segment tells you who should get your email. Smart Delivery is what we built for the next step: monitoring your deliverability metrics and data sources, so you know when to tighten or expand your Base Sending Segment, and when to cut sources that are doing more harm than good.
If you've run the prompt and want a second set of eyes on your numbers, send me a reply.
